Skip to main content
SAWC Spring 2026 Posters Main banner
Left Banner
Icon Legend
This session is not in your schedule.
This session is in your schedule. Click again to remove it.

Practice Innovations

(PI-041) Applying Semantic Artificial Intelligence to Improve Data Retrieval and Analysis in Wound Care

Bruce L K Davey, PhD
Introduction: Wound care services generate large volumes of structured entries and free text notes. Turning this information into practical insight often requires specialist analysts and hand-crafted reports. Clinicians may not know which questions their data can answer, and traditional keyword or filter-based tools struggle with inconsistent terminology across forms, sites, and documentation habits. As part of a broader wound care Artificial Intelligence (AI) initiative, we explored whether a semantic AI interface could act as a conversational “analytics assistant,” enabling wound care teams to ask clinically framed questions in natural language and receive interpretable, data-backed answers without needing database expertise. 

Methods: We designed a prototype semantic AI layer on top of an existing wound care information system configured with realistic, de-identified patient, wound, and assessment data. The interface accepts free-text clinical questions, interprets key concepts such as wound type, treatment, time frame, and outcome, and then maps them to relevant fields and measures within the underlying database. The system returns simple visual and text summariesrather than raw queries, emphasizing transparency, such as the ability to review how patients, time windows, and metrics were selected. Typical test scenarios focused on healing trajectories, documentation completeness, and treatment response in common wound types. 

Results: In this evaluation environment, the semantic interface successfully answered complex, multi-step questions such as “Which patients with venous leg ulcers showed delayed healing after compression therapy?” without manual query writing. It consistently produced patient cohorts, longitudinal views, and protocol-compliance summaries that aligned with expectations based on known test cases, and responses were generated in seconds using conversational input. 

Discussion: This work suggests that semantic AI can lower the barrier to advanced wound care analytics by hiding schema complexity, while preserving clinical transparency. A conversational assistant layered on top of existing databases may help clinicians and quality teams more quickly identify healing patterns, gaps in documentation, and high-risk groups, forming a foundation for future decision support and multi-site benchmarking.